Currently, the only way we test the Evergreen release automation is by attempting a release. We should introduce tooling to allow testing this configuration so that we know it will work when we actually try a release.

We can use --dry-run to ensure a release doesn't actually get published.